On Fri, Jun 02, 2006 at 10:15:53AM +0200, Wouter Verhelst wrote: > > So I guess if the current dip doesn't look out of the ordinary to any of the > > porters, we can wait and see. > Currently, I'm suspecting a pre-freeze upload frenzy, since there's a > dip for almost every architecture (though most other architectures > recovered within one day).
I've found (and fixed) another reason for being slow: Vivaldi was running a busy polling updatedb in one of its chroots. For some reason an upgrade triggered the start of cron in the caml chroot and was pulling CPU power away from the buildd, which runs at a lower priority. Christian found another problem with two parallel started buildds, which not only causes CPU wasted time, but as well build failures. See the other thread for this. > I'm not aware of any useful statistics about > source uploads, though (http://ftp-master.debian.org/size-quarter.png is > too convoluted to draw any serious conclusions from), so it's hard to be > sure. Well, someone could use the database on Buildd.Net to generate such a graph. I assume that the "total" number reflects the number of source packages. Combined with the number of changes in Needs-Build it might be possible to get the wanted data out of the existing data. Otherwise you'll need to track *-all.txt for changes to state Needs-Build and look for new packages/newly uploaded source packages. I'll put that on the Buildd.Net ToDo list... -- Ciao... // Fon: 0381-2744150 Ingo \X/ SIP: [EMAIL PROTECTED] gpg pubkey: http://www.juergensmann.de/ij/public_key.asc -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]